So you want to go from zero to actually shipping things with Claude Code in about an hour. No developer background needed. All you need is a working laptop and a Claude subscription. That's it. I'm Eric, and I used to be a senior software engineer at Amazon and Microsoft. Now I run my own startup, Booker.ai, and I built the entire thing with Claude Code: development, marketing, operations, you name it. I've been making videos on this for over a year, and I've seen what trips people up. This guide is the one I wish I had when I started.
Before you click away thinking this is just another Claude tutorial, here's what we're actually covering: installation, the fundamentals, file structure, the CLAUDE.md file, context, slash commands, git, connecting MCPs and CLI tools to apps like Jira and Slack, deploying what you build, and a big FAQ section pulled from my community and my own mistakes. Plus the latest updates from Anthropic on how to use Claude Code better. Let's get into it.
Installing Claude Code on Your Machine
First, you need Claude Code on your local machine. Go to the Claude documentation (just Google "Claude Code documentation") and look for the quick start. They recommend the native install. I'm on Mac, so I copy the install command they give for macOS. Open your terminal (Command + Spacebar, type "terminal", hit enter), paste the command, and run it. That's it.
Once it's installed, you have a few ways to use it. The desktop app has a toggle for Claude Code, but it's limited. You won't get all the customizations or full feature set. It's fine if you're non-technical and just want the basics. If you want real control, you can use the terminal directly, but the terminal's interface isn't great for navigating files. My recommendation is an IDE, specifically VS Code. You get the terminal inside a visual environment where you can see your files, edit them, and keep everything organized.
Setting Up VS Code for a Better Experience
Download Visual Studio Code from the official site. It'll detect your OS. Once installed, open it. A few quick shortcuts: Command + minus zooms out, Command + plus zooms in. To see your files, click the Explorer icon on the left. Right now you don't have a folder open, so click "Open Folder" and create a new one. I'll call mine "Eric Tech folder" here. That folder is where your project lives. Any file you create or delete inside VS Code reflects on your actual machine.
You can change the color theme by clicking the gear icon, then "Themes" > "Color Themes". I like Solarized Light; pick whatever you prefer. The key part is the terminal. Click the toggle panel icon at the top right (or use Control + backtick) to open the terminal at the bottom. You can resize it so you see your files above and the terminal below. That's the sweet spot for working with Claude Code.
Starting Your First Claude Code Session
In the terminal, just type claude and hit enter. That starts a Claude Code session. You'll see a prompt window where you can type. The model is shown on the status bar. Type something like "hi" and Claude will respond. It'll show you how long it took, tokens consumed, and what percentage of the context window you've used.
Now the real power: Claude Code can create and edit files in your current folder. Try asking it to create a file called "example" with some random text inside. It'll do it, and you'll see the file appear in your Explorer. You can delete files by selecting them and hitting Command + Delete, or right-click and delete. To end a session, click the trash icon in the terminal or just close it.
Understanding the File Structure and CLAUDE.md
When Claude Code works on a project, it generates a few things. You'll see a .claude folder and a CLAUDE.md file at the root of your project. The .claude folder holds internal stuff like session data. The CLAUDE.md file is where you can put instructions, context, or rules that Claude Code will follow every time it operates in that folder. Think of it as a persistent memory for your project. You can tell it your coding style, which libraries to use, or how you want it to handle errors. I'll cover that more in the advanced video, but just knowing it exists is huge.
Using Context, Slash Commands, and Git
Claude Code has a concept of context. The more you chat, the more it remembers within that session, but once you hit the context limit, it might start forgetting earlier instructions. That's why you want to keep sessions focused and use the CLAUDE.md file for things that should always be remembered.
Slash commands let you do things like /clear to reset the conversation or /compact to summarize the context and free up space. You can also use /git to manage version control directly from Claude Code. It can commit, push, create branches: all the git stuff you'd normally do in the terminal. If you're building anything serious, version control is non-negotiable, and Claude Code makes it easy.
Connecting MCPs and CLI Tools
MCPs (Model Context Protocols) and CLI tools are how you hook Claude Code into the other apps you use daily. Want it to pull tasks from Jira? There's an MCP for that. Want it to send a Slack message when a build finishes? You can set that up. The idea is to let Claude Code interact with your actual workflow, not just code in isolation. In the video I show how to configure these connections so that your agent can, for example, read a Jira ticket and then start working on the related code. It's one of the features that makes Claude Code feel less like a toy and more like a real teammate.
Deploying Your Application
Once you've built something, you need to get it live. Claude Code can help with deployment too. You can instruct it to package your app, set up environment variables, and push to a hosting service. The exact steps depend on your stack, but the process is similar to how you'd do it manually, except Claude handles the commands. I walk through a basic deployment in the video, and the key takeaway is that you don't need to leave your terminal: Claude Code can orchestrate the whole thing.
Key Takeaways
- You don't need to be a developer to start. A laptop and a Claude subscription are enough.
- Use VS Code as your IDE: it gives you a visual file explorer and a terminal side by side.
- Start a Claude Code session by typing
claudein the terminal. It can read, write, and delete files in your project folder. - The
CLAUDE.mdfile is your project's persistent instruction manual for Claude. - Slash commands like
/clearand/compacthelp manage context;/githandles version control. - MCPs and CLI tools let Claude Code connect to Jira, Slack, and other apps you already use.
- The $20 plan is fine for learning, but daily heavy use will hit limits fast. Combine it with local models or Open Router (using models like DeepSeek V3) to stay within budget.
- The fundamentals taught here won't be outdated in a month. I've been teaching Claude Code for over a year, and the core concepts remain the same even as new models drop.
FAQ: The Questions Everyone Asks
Is the $20 plan enough?
If you're just starting, yes. But if you're using Claude Code every day, I'll be upfront: you will hit the rate limits. The $20 plan runs out fast. To stay under budget, you have two options. One, use a local model if your machine is powerful enough. Two, connect Claude Code to Open Router and use cheaper models like DeepSeek V3 as a substitute for Sonnet, saving Opus for the really tough tasks. I have a video on saving tokens, and another one coming soon on skills that reduce usage.
How do I stop losing context?
Watch my video on controlling context. The short answer: use /compact to summarize, keep sessions focused, and put long-term instructions in CLAUDE.md.
Is it safe to run on my machine?
Yes, if you set the right permissions. In the video I cover the four permission modes. Pick the one that gives Claude Code enough access to do its job without letting it run wild. Always review what it's about to do before approving.
Do I need Cursor to use Claude Code?
No. You can use VS Code, Cursor, or any IDE with a terminal. Cursor is an IDE with AI built in, but you don't need it. For beginners, VS Code is the simplest and most minimal option.
Won't all this be outdated in a month?
No. The foundation stays the same. I've been making Claude Code videos since it launched, and while the models have evolved (Sonnet, Opus, now Opus 4.8), the way you install, structure projects, use context, and manage permissions hasn't changed. You need these fundamentals before you can tackle advanced stuff like agent loops or dynamic workflows.
How do you handle version control?
Use the /git slash command. Claude Code can commit, push, and manage branches for you. It's built in.
Can I use Claude Code for free?
You need a subscription, but you can reduce costs by mixing in free or cheap models via Open Router. I have a video on running Claude Code for free that goes deeper.
What if I want to go further?
Inside my school community, Eric Tech AI Builders, I have a full course on mastering Claude Code, building AI agents, automations, and shipping real SaaS products. We do weekly live calls and Q&As. And my next video will cover intermediate topics: agents, skills, spec-driven development, knowledge graphs, and local AI. If that sounds interesting, like this video and subscribe.
So that's the full beginner-to-pro path in one shot. The tools are there, the fundamentals are stable, and you can start building today. If you found this helpful, drop a comment, and I'll see you in the next one.